Consider any given action, X, performed by any agent S. Consider a description of X: such a description will have a largely causal story (S X’d), while an explanation will feature a belief/desire story, and most significantly for our purposes here, a reasons story (though the causal story will be significant as well). The overall point here will be to show a couple of things: that reasons figure prominently in a rational explanation of action but in a non-causal way, reasons point to a conception of freedom that is positive, as opposed to negative (being free from causal forces, for example), and that the causal and reasons stories are distinct but not opposed to each other – an agent can be caused to have a reason to X, in other words, without the causal story being the figuring more prominently than the reasons story.
Why can’t reasons serve as causes? More pointedly: why, if reasons can’t serve as causes, are they exempt from the causal order? The answer to this is fairly simple: causal relations occur between events, and reasons aren’t events. Therefore, reasons can’t enter into causal relations. Now we have to distinguish quite carefully here just what we’re trying to do with these reasons and causes: we want to rationally explain S’s X-ing, and a purely causal story or even a generally causal story simply isn’t up to ontological snuff to give us that explanation. Clearly the causal story is important, but rational explanation doesn’t appeal to events to explain but appeals to reasons, which as noted above, are not events. What are they? Reasons are, in a nutshell, what our mental states are about. This links them closely to propositional attitudes – indeed, Davidson took reasons to be composed largely of pro-attitudes and built his ‘reasons as causes’ theory of action largely on that point. Davidson saw sentences such as ‘X did Y because’ as requiring a causal answer in large part due to the ‘because’: if what came after ‘because’ wasn’t causal, Davidson believed we were left without an adequate analysis of Y. Scott Sehon notes that a fully adequate answer was in plain view:
Davidson asks for an alternative to a causal construal of the ‘because’ in action explanation…it seems that a telelogical construal should have been obvious: we can read “Vera went to the cafe because she wanted coffee” as claiming that Vera went to the cafe in order to satisfy her desire for coffee. (Telelogical Realism: Mind, Agency and Explanation, p. 156)
One wonders if there aren’t conceptual confusions afoot here. Wittgenstein, deep in the Blue Book, picks up on some of the conceptual issues at play here in the reasons/causes dialectic; perhaps unsurprisingly, he takes the issue to be primarily grammatical (Wittgenstein also seems to equate ‘reason’ with ‘motive’). Causes, Wittgenstein argues, can’t be known but can only be conjectured. Causes are a hypothesis, and the tangle occurs because we take the question of ‘why X’ to be able to be answered by both a causal and a motive/reason story. This happens because we take ourselves to know our motives, we also take it that we can know the causes, because, on this scheme, ‘motives are a cause of which we are aware’ (The Blue Book). The confusion, then, is this: we take motives/reasons and causes to both answer the question of ‘why’ because we mistakenly take them to be species of each other. Perhaps this helps explain how reasons came to be seen as causes?
This to me seems to anticipate – in spirit anyway – Sellars concepts of the ‘space of reasons’ and the ‘space of nature/natural law/law’. So in the explanatory story of why you X, there’s the causal element – someone pushed your daughter into the pool (call this the space of causes). The flip side of this is the space of reasons: this is the normative space in which we make judgements, committments, are bound by conceptual norms, are liable to assesments. These two spaces are distinct but not opposed: there can be causal antecedents for an agent having a reason without that agent losing any freedom, because on this account freedom is the capacity to endorse internal reasons, or, to put it another way, to adopt normative statuses. Your daughter’s being pushed into the pool is a causal event, but your acting according to norms is a free event, not because you are free from any sort of causality but because you are acting within the space of reasons, which is ‘rational freedom’.